Ask a Neuroscientist: Auditory Hallucinations

I’ve noticed that when I’m sleep-deprived I often hear music, usually a song I’ve listened to recently. It’s not a hallucination exactly, in that I can tell that it’s not real. But it’s so vivid that I can hear details of production or small sound effects that I never noticed before - and which are actually in the song. It’s not unpleasant exactly, sometimes it’s even kind of enjoyable, though it can be distracting. My question is: is it possible for some parts of the brain to be stimulated by lack of sleep (i.e. the aural imagination parts), even while others are impaired?
— Mischa

Hi Mischa:

Why can we hear music that isn’t really there? This question gnaws at some fairly complex and incompletely understood neuroscience phenomena. A complete answer would require us to understand and explore consciousness, how we sense the world around us, and even how we discern the real from the unreal. Despite vast advances in neuroscience knowledge within the last century or so, we are still far from understanding these complex processes. Luckily, the musical experiences that you describe actually sound like they might be related to two recognized and relatively well-studied conditions. As such, these conditions provide us with a framework for understanding how we can perceive music that isn’t really there.

The two conditions in question are musical hallucinosis and hypnagogic hallucinations. People with musical hallucinosis report hearing complex musical sounds in the absence of an objectively verifiable sound source. This condition is rare and usually accompanies hearing loss. Hypnagogic hallucinations, on the other hand, describe any hallucination that occurs during the transition between wakefulness and sleep. This transitional period is also associated with decreased perception of bodily sensations. In fact, at the core of hallucinations lies a breakdown in how signals from the outside world are translated into experience. As I will explain, you are quite correct in thinking that hallucinations are, in a sense, associated with greater brain activity.

For our purposes, it is useful to think of awareness and hallucination as overlapping rather than distinct phenomena. Put simply, awareness is the sensory perception of an event that you and I can both see/hear/etc. You and I sense this event and we can both verify that it exists in the real world — you see a red rose and I see a red rose, you hear Blur and I hear Blur. Conversely, a hallucination is something that only one of us experiences in a real and tangible way — I see a dagger before me but you do not, then while pondering weak and weary I hear a raven quoth “nevermore”, but you do not hear the raven. And yet, to me, the quotation by the ominous bird of yore seems just as real during the hallucination as when it is verifiably produced by an individual Corvus covax and heard by everyone around me. So by extension, during a musical hallucination I hear a song that no one else hears. I am convinced that the sound comes from some source outside of me and that it is not simply imagined. To the person having the experience, the real and the hallucinated appear to be one and the same thing.

Hallucinations are possible because our brains are not mere conveyors, but rather interpreters of information. One of the major tasks performed by our brains is to use the information received from our sensory organs to create for us an experience of our world. In fact, if we could somehow perceive the world without interpretation, then hallucinations probably could not exist. Think of your experience of the world as being more like Microsoft Windows rather than the software code running in the background; we are only privy to the interpretation and not to the raw data. Of course, every experience of the world — from the most mundane touch of the computer mouse to the sound of the most beautiful piece of music — starts as fairly ordinary signals that are then subjected to very complex processing. So, anything that disrupts any part of the input pathways or the interpretation systems can lead to changes in perception and even hallucinations. So, let’s consider music. The violin string becomes a sound wave of a specific frequency, this reaches the ear making little bones vibrate which make the eardrum vibrate, which make little hair cells in the ear vibrate, sending an electrical signal along the auditory (hearing) nerve to the brain, which processes the signals, considering the signals in the context of what was heard before, whether they are pleasurable, and finally determines whether it is music. The brain will lay down a small memory of the tune and any associated emotions; then, the more the tune is heard, the stronger becomes the memory of the pitch, timber, and tempo that make up the musical piece. And the next time the brain receives these same signals, you will recall the music; perhaps you might even recognize that sound as a specific musical piece as soon as you hear the first few notes. Thus, in a very important way, listening to music involves pattern matching and recall.

By matching incoming sounds to known musical patterns, the brain cuts down on having to process every piece of information, every single time.  The brain does not work like a blank slate, taking in information, and recreating the world every time. The brain makes predictions about the world and uses input from the world outside to support or reject its predictions. Incoming information is used to verify the predictions. We’ve all consciously experienced this prediction system breaking down when listening to music. Did you ever listen to a song thinking “I know that tune” but it turned out to be something else? Maybe you even noticed a moment of discordance when your expectation did not match up with the reality? In the case of the song, your brain received a few familiar notes — it turned on the “ice ice baby” program to avoid having to process the whole thing all over again. However, having then received more information, it decided that the original prediction was wrong (in fact the song was “under pressure” by Queen and David Bowie). Such shortcuts decrease the brain’s processing requirements, but they also make it likely that we can be wrong. And when those predictions are very wrong and unrestrained by external information, we can hallucinate. That the brain really can create experiences of the world is demonstrated to us every night when we have a dream. But it can also been shown experimentally during states of complete wakefulness. In 2009, researchers in London wondered what would happen if people were temporarily deprived of all vision and hearing. They placed 19 people into an anechoic chamber, which is built to be essentially completely sound- and light-proof. In this environment, completely devoid of their usual sensory input, completely healthy people reported seeing various visions such as faces and shapes. For an amusing version of this phenomenon, watch episode 16 in the 10th season of the Simpsons, where Homer and Lisa spend time hallucinating in a sensory deprivation chamber. What multiple sensory deprivation studies (and the Simpsons episode) tell us, is that even the temporary deprivation of sight and sound input can lead to hallucinations. What they also tell us, is that sensory input is necessary to keep the brain’s creations from getting too out of hand.

So, now imagine that the brain creates the “ice ice baby” experience but there is no feedback to stop the program! It could be because there is nothing to correct the brain’s incorrect prediction (e.g. in the case of hearing loss). Or, it could be because the normal process of verification breaks down (e.g. during the transition to and from sleep). Might the experience continue? In fact, this is what appears to happen during musical hallucinosis and hypnagogic hallucinations. The input from our senses actually grounds us and sets up boundaries for our experience. Anything that alters the sensation or its interpretation can alter our experience by decreasing the feedback to the brain’s fanciful creations. As it happens, this is exactly what happens when one is falling asleep, during sleep, and also in the presence of various nervous system disorders, psychiatric conditions, and after the ingestion of drugs from the hallucinogen class. Musical hallucinations can take place when the spontaneously generated brain activity or predictions are not properly constrained by feedback from the world. This is why musical hallucinosis usually manifests when people lose their hearing. In blindness, there is a similar condition involving vision, called the Charles Bonnet Syndrome. In both conditions, the normal sensory input is lacking and the brain is allowed to run relatively unrestrained. In fact imaging and electroencephalogram (EEG) studies suggest that brain activation during hallucinations and other abnormal states of perception is similar to that seen during normal perception [see: here, here, or here, and especially here and here].  So, in a sense, when you experience musical hallucinations, there is extra activation of the brain. This is because the brain activates and produces an experience of music at times when it should not.

While the exact mechanisms involved in the generation of hallucinations are not understood, there is a model that fits with our current understanding of the brain. In a sense, the world of musical hallucinations is one that is dominated by memories and predictions that are untempered by reality. And because the brain is creating the experience all on its own, it is possible that it highlights certain features of the music that are normally inhibited — such as sound production artefacts (e.g. performer breaths, violin bow tapping, audible pops, etc). While I cannot definitively confirm the exact biological mechanisms of your specific symptoms, I hope that I shed a little light on some of the possible mechanisms at play.